Monitoring Multiple Greenplum Database Clusters

The Greenplum Command Center Multi-cluster view displays health status for multiple Greenplum Database clusters. The charts can be divided into categories.

Each Greenplum Database cluster that appears in the Multi-cluster view must have its own Command Center instance. The Multi-cluster view is hosted on a single, designated master Command Center instance. The master instance retrieves the health status data from each of the clusters’ Command Center instances.

Clicking the status panel for a cluster loads that cluster’s Command Center into a new browser window. If auto-login is enabled in the multi-cluster configuration file, the login screen is bypassed and the Dashboard is displayed.

Multi-cluster support is enabled by deploying a configuration file to the Command Center instance directories. See Enabling Multi-Cluster Support for instructions to set up the Multi-cluster page.

By default, any Command Center user may view multi-cluster status. Access to clusters can be restricted in the multi-cluster configuration file.

The Multi-cluster view displays the following information for each cluster.

Cluster Name
The name for the cluster, as specified in the multi-cluster configuration file.

Database State
The current state of the Greenplum Database cluster. The state can be one of the following:

  • Normal: The database is functioning with no major errors or performance issues.
  • Segment(s) Down: The database is in change-tracking mode or resync mode. Overall performance and system reliability is greatly reduced. See the Pivotal Greenplum Database System Administrator Guide for information about resolving this condition.
  • Database Unreachable: The Greenplum Performance Monitor agent cannot connect to the database. The database is likely down. See the Pivotal Greenplum Database System Administrator Guide for troubleshooting information.
  • Unbalanced: Some segments are not running in their preferred roles. That is, primaries are running as mirrors and mirrors are running as primaries, resulting in unbalanced processing.
  • Resyncing: The database is performing a recoverty or rebalance operation.

The elapsed time since the Greenplum Database system was last started.

GPDB Version
The version of the Greenplum Database software each monitored cluster is running.

The number of active Greenplum Database sessions (client connections).

Active Queries
The number queries queued or currently executing in the database.